Remote Measurement and Detection System for Electric Vehicle Charging Process and Its Application

Aiming at the problems of insufficient multi-dimensional data fusion, real-time response delay and data privacy protection in the charging process of electric vehicles, this research proposes a remote measurement and control system architecture that integrates edge computing and cloud intelligence. By constructing a three-level collaborative system of “perception edge cloud”, a high-precision sensor array is designed to achieve synchronous collection of charging voltage (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$0-1000 ~\mathrm{V}$</tex>, error <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\leqslant 1.5 \%$</tex>), current (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$0-250 ~\mathrm{A}$</tex>, error <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\leqslant 1.2 \%$</tex>), charging messages, and environmental parameters; The edge computing unit is developed based on the heterogeneous architecture of ARM+FPGA, and the fuzzy neural network is integrated to realize rapid discrimination of local abnormal states (response time <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$&lt;150 ~\text{ms}$</tex>); Build a cloud based intelligent platform that uses quantile regression models and federated learning techniques to trace fault chains and predict battery degradation, ensuring data privacy and security. System testing shows that the multi-dimensional data spatiotemporal alignment error is controlled within 1 ms, the overcurrent fault detection delay is 120 ms, the accuracy of anomaly detection is <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{9 5. 3 \%}$</tex>, the cloud platform supports concurrent access of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{1 0 0}$</tex> charging piles, the peak processing capacity is <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$4800 \text{eq} / \mathrm{s}$</tex>, and the operation and maintenance response efficiency is improved by more than 70 % compared to traditional solutions. The research results break through the limitations of single parameter monitoring and provide engineering solutions for the safe operation of large-scale charging networks. The experiment has verified the effectiveness of edge cloud collaborative architecture in real-time, reliability, and data security. In the future, we will focus on the design of distributed cloud architecture, optimization of transfer learning models, and improvement of adaptability to complex environments, promoting the evolution of the system towards intelligence and universality.